Re: orgstruct-mode with custom headline prefix

[Christopher Schmidt <christopher@ch.ristopher.com>]

Bastien <bzg@altern.org> writes:

Hi Bastien,

>> Bastien <bzg@altern.org> writes:
>>> Can you give an indication on how it impacts the generation of
>>> agenda?  This is a potential blocker.
>>
>> It should not impact agenda generation at all - neither feature nor
>> performance-wise.  All non-trivial changes are in parts of the code
>> that directly relate to orgstruct-mode or
>> org-\(forward\|backward\)-heading-same-level.  These functions and
>> their internals are not used non-interactively.
>
> org-outline-regexp-bol, org-\(forward\|backward\)-heading-same-level
> are fundamental vars/functions -- any change here is likely to impact
> the agenda generation.  So while I hope you're right, I'd feel better
> with figures and tests :)

org-outline-regexp-bol value is ^org-outline-regexp.  There is one place
in org where org-outline-regexp-bol is actually set -
org-with-limited-levels.  It is let-bound to (concat "^"
org-outline-regexp).

BTW what is the reason for having both org-outline-regexp and
org-outline-regexp-bol?  Matching the latter should be
faster than the former.

org-\(forward\|backward\)-heading-same-level are referenced
non-interactively in three functions - org-forward-element,
org-backward-element and org-copy-subtree.  I checked the implementation
of these functions and I did not see how the new behaviour should break
these functions.

After all, the functions now do what they are supposed to do.  Try this:

    * Org
    ** Achim
    ** Bastien
    ** Carsten
    * Diary

Place point on Achim and C-c C-b.  Place point on Casten and C-c C-f.

All my agendas generate just fine.  I am not exactly sure how I can
provide figures and tests.  We do not have any agenda test cases in the
first place.

Please take a look at the patch when you have time.  The diff is not
that scary.

Greetings,

        Christopher